In [18]:
import pandas as pd
import numpy as np
from sklearn.metrics import roc_auc_score, accuracy_score, precision_score, recall_score, f1_score
from sklearn.linear_model import LogisticRegression
from sklearn.tree import DecisionTreeClassifier
from sklearn.svm import SVC
from sklearn.ensemble import RandomForestClassifier
from sklearn.discriminant_analysis import LinearDiscriminantAnalysis, QuadraticDiscriminantAnalysis
from sklearn.cluster import KMeans
from sklearn.neighbors import KNeighborsClassifier
from sklearn.naive_bayes import GaussianNB
from sklearn.metrics import precision_recall_fscore_support
from sklearn.metrics import mean_squared_error as mse
from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import cross_validate
from sklearn.model_selection import train_test_split
from sklearn.ensemble import ExtraTreesClassifier
from sklearn.feature_selection import SelectFromModel
from sklearn.svm import LinearSVC
from sklearn.ensemble import VotingClassifier
from sklearn.feature_selection import RFECV
In [19]:
from IPython.core.interactiveshell import InteractiveShell
InteractiveShell.ast_node_interactivity = "all"
In [20]:
import warnings
warnings.filterwarnings('ignore')
In [21]:
initial_data = pd.read_csv('/Users/MichaelMiao/Documents/career/Jupyter_files/Python/Yuyan/train_hot.csv',index_col=None)
In [23]:
initial_data.head()
Out[23]:
In [24]:
initial_data=initial_data.iloc[0:,1:]
In [25]:
initial_data['HTWins']=initial_data['HTWins'].map({'Yes':1,'No':0})
In [26]:
initial_data.head(3)
Out[26]:
In [27]:
tester= pd.read_csv('/Users/MichaelMiao/Documents/career/Jupyter_files/Python/Yuyan/test_hot.csv',index_col=None)
In [28]:
tester.head(3)
tester=tester.iloc[0:,1:]
Out[28]:
In [29]:
tester.head(3)
Out[29]:
In [ ]:
In [30]:
initial_data.drop(initial_data[['date']], axis=1, inplace=True)
initial_data.drop(initial_data[['id']], axis=1, inplace=True)
initial_data.drop(initial_data[['gameID']], axis=1, inplace=True)
initial_data.head()
tester.drop(tester[['date']], axis=1, inplace=True)
tester.drop(tester[['id']], axis=1, inplace=True)
tester.drop(tester[['gameID']], axis=1, inplace=True)
tester.head()
Out[30]:
Out[30]:
In [31]:
tester.head()
initial_data.head()
Out[31]:
Out[31]:
In [32]:
import seaborn as sns
import matplotlib.pyplot as plt
sns.countplot(initial_data['HTWins'],label="Sum")
plt.show()
Out[32]:
In [33]:
X = initial_data.drop(['HTWins'],axis=1)
X.head()
Out[33]:
In [34]:
y = initial_data[['HTWins']]
y.head(3)
Out[34]:
In [35]:
X_train, X_test, y_train, y_test = train_test_split(X,y,test_size=0.34)
In [36]:
LR = LogisticRegression()
scoring = ['accuracy', 'precision_macro', 'recall_macro' , 'f1_weighted', 'roc_auc']
scores = cross_validate(LR, X_train, y_train, scoring=scoring, cv=20)
sorted(scores.keys())
LR_fit_time = scores['fit_time'].mean()
LR_score_time = scores['score_time'].mean()
LR_accuracy = scores['test_accuracy'].mean()
LR_precision = scores['test_precision_macro'].mean()
LR_recall = scores['test_recall_macro'].mean()
LR_f1 = scores['test_f1_weighted'].mean()
LR_roc = scores['test_roc_auc'].mean()
Out[36]:
In [37]:
LR_accuracy
Out[37]:
In [45]:
decision_tree = DecisionTreeClassifier()
scoring = ['accuracy', 'precision_macro', 'recall_macro' , 'f1_weighted', 'roc_auc']
scores = cross_validate(decision_tree, X_train, y_train, scoring=scoring, cv=20)
sorted(scores.keys())
dtree_fit_time = scores['fit_time'].mean()
dtree_score_time = scores['score_time'].mean()
dtree_accuracy = scores['test_accuracy'].mean()
dtree_precision = scores['test_precision_macro'].mean()
dtree_recall = scores['test_recall_macro'].mean()
dtree_f1 = scores['test_f1_weighted'].mean()
dtree_roc = scores['test_roc_auc'].mean()
Out[45]:
In [47]:
dtree_accuracy
Out[47]:
In [48]:
SVM = SVC(probability = True)
scoring = ['accuracy','precision_macro', 'recall_macro' , 'f1_weighted', 'roc_auc']
scores = cross_validate(SVM, X_train, y_train, scoring=scoring, cv=2)
sorted(scores.keys())
SVM_fit_time = scores['fit_time'].mean()
SVM_score_time = scores['score_time'].mean()
SVM_accuracy = scores['test_accuracy'].mean()
SVM_precision = scores['test_precision_macro'].mean()
SVM_recall = scores['test_recall_macro'].mean()
SVM_f1 = scores['test_f1_weighted'].mean()
SVM_roc = scores['test_roc_auc'].mean()
Out[48]:
In [23]:
SVM_accuracy
Out[23]:
In [38]:
LDA = LinearDiscriminantAnalysis()
scoring = ['accuracy', 'precision_macro', 'recall_macro' , 'f1_weighted', 'roc_auc']
scores = cross_validate(LDA, X_train, y_train, scoring=scoring, cv=20)
sorted(scores.keys())
LDA_fit_time = scores['fit_time'].mean()
LDA_score_time = scores['score_time'].mean()
LDA_accuracy = scores['test_accuracy'].mean()
LDA_precision = scores['test_precision_macro'].mean()
LDA_recall = scores['test_recall_macro'].mean()
LDA_f1 = scores['test_f1_weighted'].mean()
LDA_roc = scores['test_roc_auc'].mean()
Out[38]:
In [39]:
LDA_accuracy
Out[39]:
In [51]:
QDA = QuadraticDiscriminantAnalysis()
scoring = ['accuracy', 'precision_macro', 'recall_macro' , 'f1_weighted', 'roc_auc']
scores = cross_validate(QDA, X_train, y_train, scoring=scoring, cv=20)
sorted(scores.keys())
QDA_fit_time = scores['fit_time'].mean()
QDA_score_time = scores['score_time'].mean()
QDA_accuracy = scores['test_accuracy'].mean()
QDA_precision = scores['test_precision_macro'].mean()
QDA_recall = scores['test_recall_macro'].mean()
QDA_f1 = scores['test_f1_weighted'].mean()
QDA_roc = scores['test_roc_auc'].mean()
Out[51]:
In [52]:
QDA_accuracy
Out[52]:
In [53]:
random_forest = RandomForestClassifier()
scoring = ['accuracy', 'precision_macro', 'recall_macro' , 'f1_weighted', 'roc_auc']
scores = cross_validate(random_forest, X_train, y_train, scoring=scoring, cv=20)
sorted(scores.keys())
forest_fit_time = scores['fit_time'].mean()
forest_score_time = scores['score_time'].mean()
forest_accuracy = scores['test_accuracy'].mean()
forest_precision = scores['test_precision_macro'].mean()
forest_recall = scores['test_recall_macro'].mean()
forest_f1 = scores['test_f1_weighted'].mean()
forest_roc = scores['test_roc_auc'].mean()
Out[53]:
In [54]:
forest_accuracy
Out[54]:
In [30]:
KNN = KNeighborsClassifier()
scoring = ['accuracy', 'precision_macro', 'recall_macro' , 'f1_weighted', 'roc_auc']
scores = cross_validate(KNN, X_train, y_train, scoring=scoring, cv=20)
sorted(scores.keys())
KNN_fit_time = scores['fit_time'].mean()
KNN_score_time = scores['score_time'].mean()
KNN_accuracy = scores['test_accuracy'].mean()
KNN_precision = scores['test_precision_macro'].mean()
KNN_recall = scores['test_recall_macro'].mean()
KNN_f1 = scores['test_f1_weighted'].mean()
KNN_roc = scores['test_roc_auc'].mean()
Out[30]:
In [31]:
KNN_accuracy
Out[31]:
In [32]:
bayes = GaussianNB()
scoring = ['accuracy', 'precision_macro', 'recall_macro' , 'f1_weighted', 'roc_auc']
scores = cross_validate(bayes, X_train, y_train, scoring=scoring, cv=20)
sorted(scores.keys())
bayes_fit_time = scores['fit_time'].mean()
bayes_score_time = scores['score_time'].mean()
bayes_accuracy = scores['test_accuracy'].mean()
bayes_precision = scores['test_precision_macro'].mean()
bayes_recall = scores['test_recall_macro'].mean()
bayes_f1 = scores['test_f1_weighted'].mean()
bayes_roc = scores['test_roc_auc'].mean()
Out[32]:
In [33]:
bayes_accuracy
Out[33]:
In [38]:
models_initial = pd.DataFrame({
'Model' : ['Logistic Regression', 'Decision Tree', 'Support Vector Machine', 'Linear Discriminant Analysis', 'Quadratic Discriminant Analysis', 'Random Forest', 'K-Nearest Neighbors', 'Bayes'],
'Fitting time': [LR_fit_time, dtree_fit_time, SVM_fit_time, LDA_fit_time, QDA_fit_time, forest_fit_time, KNN_fit_time, bayes_fit_time],
'Scoring time': [LR_score_time, dtree_score_time, SVM_score_time, LDA_score_time, QDA_score_time, forest_score_time, KNN_score_time, bayes_score_time],
'Accuracy' : [LR_accuracy, dtree_accuracy, SVM_accuracy, LDA_accuracy, QDA_accuracy, forest_accuracy, KNN_accuracy, bayes_accuracy],
'Precision' : [LR_precision, dtree_precision, SVM_precision, LDA_precision, QDA_precision, forest_precision, KNN_precision, bayes_precision],
'Recall' : [LR_recall, dtree_recall, SVM_recall, LDA_recall, QDA_recall, forest_recall, KNN_recall, bayes_recall],
'F1_score' : [LR_f1, dtree_f1, SVM_f1, LDA_f1, QDA_f1, forest_f1, KNN_f1, bayes_f1],
'AUC_ROC' : [LR_roc, dtree_roc, SVM_roc, LDA_roc, QDA_roc, forest_roc, KNN_roc, bayes_roc],
}, columns = ['Model', 'Fitting time', 'Scoring time', 'Accuracy', 'Precision', 'Recall', 'F1_score', 'AUC_ROC'])
models_initial.sort_values(by='Accuracy', ascending=False)
Out[38]:
In [45]:
correlation = initial_data.corr()
mask = np.zeros_like(correlation, dtype=np.bool)
mask[np.triu_indices_from(mask)] = True
f, ax = plt.subplots(figsize=(100, 100))
cmap = sns.diverging_palette(180, 20, as_cmap=True)
sns.heatmap(correlation, mask=mask, cmap=cmap, vmax=1, vmin =-1, center=0,
square=True, linewidths=.5, cbar_kws={"shrink": .5}, annot=True)
plt.show()
Out[45]:
